Seasonal Performance of a Nonhydrostatic Global Atmospheric Model on a Cubed-Sphere Grid

Abstract

The Korean Integrated Model (KIM), a recently developed nonhydrostatic global atmospheric model over a cubed-sphere grid, was deployed in April 2020 as an operational weather forecasting model. As its application extends to research and predictions longer than the weather time scale, we evaluated the ability of the KIM on seasonal ensemble simulation for the boreal winter and summer cases with respect to seasonal mean biases. The results are compared with those obtained from a conventional hydrostatic spectral model, which has been widely used for seasonal simulations and in climate research. To isolate the origin of the error sources, the same physics packages is used in both the KIM and the reference models. The simulated mean states are very close to the reanalysis for the selected cases. Most large-scale fields from the KIM are comparable to those from the reference model, which implies that the general features of large-scale variables and precipitation are highly governed by physical parameterizations, and that the physics-dynamics coupling is stable in a long-term simulation. Large-scale tropical circulations, such as the Hadley and Walker circulations, need to be improved for applications related to future changes and climate projections. Moreover, the results reveal that the simulated global precipitation band is misplaced and the heat fluxes over oceans are relatively misrepresented near the eastern boundaries of tropical and subtropical regions. This analysis suggests the necessity of realistic atmosphere-ocean interactions that reflect ocean overturning circulation via ocean coupling as well as the refinement of deep and shallow convection schemes. © 2021. The Authors. Earth and Space Science published by Wiley Periodicals LLC on behalf of American Geophysical Union.